3 ms·
Many people still believe that LLMs can only reproduce what is in their training set.
by WarmWash 1mo ago
Many people still believe that LLMs can only reproduce what is in their training set.
- stymaar 1mo agoThat'd be too restrictive, but to get improvement in a particular domain you definitely need to train specifically for it. Just cramming more random internet text in a bigger model has stopped being a effective way of scaling since at least mid 2023.
- maleldil 1mo agoNo, you don't. There have been technological advances beyond just "more random Internet text", and they lead to more powerful models with more refined emergent behaviours.